Vancouver Weather Analysis
Intro
Basic analysis of Vancouver weather patterns:
- from Govt. of Canada Historical Weather Data website
- monthly
- precipitation, temperature
Data Collection
Data is collected from Govt of Canada website:
- search: http://climate.weather.gc.ca/historical_data/search_historic_data_e.html
- select station of interest by searching for stations with ‘vancouver’ at start of name
- select ‘daily’, select ‘year’ (note that year shown is end of period data is available for; most stations have very little data)
- ‘Go’ goes to a page to access results for year
(process is automated in ‘van-weather-import.R’)
## [1] "Winter" "Spring" "Summer" "Fall"Earliest date: 1970-01-01
Most recent date: 2021-07-02
` 
Check data structure and summary:
## spec_tbl_df [18,810 × 11] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
## $ Date : Date[1:18810], format: "1970-01-01" "1970-01-02" ...
## $ Year : Factor w/ 52 levels "1970","1971",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ Month : Factor w/ 12 levels "1","2","3","4",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ Day : num [1:18810] 1 2 3 4 5 6 7 8 9 10 ...
## $ Season : Factor w/ 4 levels "Winter","Spring",..: 1 1 1 1 1 1 1 1 1 1 ...
## $ Season.Yr : num [1:18810] 1970 1970 1970 1970 1970 1970 1970 1970 1970 1970 ...
## $ Max.Temp : num [1:18810] 6.1 2.2 5 2.8 1.7 5.6 3.3 4.4 7.2 5.6 ...
## $ Min.Temp : num [1:18810] -1.7 0 -1.7 -3.3 -2.8 -1.7 -3.3 0 2.8 -2.8 ...
## $ Mean.Temp : num [1:18810] 2.2 1.1 1.7 -0.3 -0.6 2 0 2.2 5 1.4 ...
## $ Total.Precip: num [1:18810] 0 1 0.5 0 0 0 0 4.6 6.6 0.5 ...
## $ Station : chr [1:18810] "VANCOUVER INT'L A" "VANCOUVER INT'L A" "VANCOUVER INT'L A" "VANCOUVER INT'L A" ...
## - attr(*, "spec")=
## .. cols(
## .. Date = col_date(format = ""),
## .. Year = col_double(),
## .. Month = col_double(),
## .. Day = col_double(),
## .. Season = col_character(),
## .. Season.Yr = col_double(),
## .. Max.Temp = col_double(),
## .. Min.Temp = col_double(),
## .. Mean.Temp = col_double(),
## .. Total.Precip = col_double(),
## .. Station = col_character()
## .. )## Date Year Month Day
## Min. :1970-01-01 1972 : 366 1 :1612 Min. : 1.00
## 1st Qu.:1982-11-16 1976 : 366 3 :1612 1st Qu.: 8.00
## Median :1995-10-02 1980 : 366 5 :1612 Median :16.00
## Mean :1995-10-02 1984 : 366 7 :1583 Mean :15.73
## 3rd Qu.:2008-08-16 1988 : 366 8 :1581 3rd Qu.:23.00
## Max. :2021-07-02 1992 : 366 10 :1581 Max. :31.00
## (Other):16614 (Other):9229
## Season Season.Yr Max.Temp Min.Temp
## Winter:4661 Min. :1970 Min. :-8.70 Min. :-15.200
## Spring:4784 1st Qu.:1982 1st Qu.: 8.90 1st Qu.: 2.800
## Summer:4724 Median :1995 Median :13.30 Median : 6.800
## Fall :4641 Mean :1995 Mean :13.83 Mean : 6.699
## 3rd Qu.:2008 3rd Qu.:19.00 3rd Qu.: 11.300
## Max. :2021 Max. :34.40 Max. : 22.400
## NA's :67 NA's :60
## Mean.Temp Total.Precip Station
## Min. :-11.50 Min. : 0.000 Length:18810
## 1st Qu.: 6.00 1st Qu.: 0.000 Class :character
## Median : 10.00 Median : 0.000 Mode :character
## Mean : 10.29 Mean : 3.214
## 3rd Qu.: 15.10 3rd Qu.: 3.600
## Max. : 28.40 Max. :91.600
## NA's :69 NA's :54- Looks like quite a few NAs for temp and precip; as long as they are spread out, shouldn’t be a problem
Check Data Relationships
- uses PerformanceAnalytics package
- Looks like temperatures are highly correlated and precipation has weak negative correlation with temperate (colder weather slightly more associated with rain)
Precipitation
Monthly precipitation data: Averages
Data by Month - ranked in order of highest precipitation
Monthly Preciptiation: Ranges
Current Month Precipitation
Monthly Comparison YoY
For each month, what has been the pattern in precipitation over the years?
Seasonal Precipitation
Season definitions:
- Winter: Dec, Jan, Feb
- Spring: Mar, Apr, May
- Summer: Jun, Jul, Aug
- Fall: Sep, Oct, Nov
Note: the year for a winter season is applied to year at end of season. So winter from Dec 2018 to Feb 2019 is considered winter of 2019.
Seasonal Precipitation Annual Ranking
- Need separate charts because each season ranked individually
- Line = median
Temperature
Check for missing data
- Something to watch for if there are suspicious results for years with higher NAs.
- Probbaly not going to be enough cause to throw whole year out. Could do some imputation of missing values, such as using averages, but probably not worth it.
Ave. Annual Temperature
- 1985 looks like an outlier, although doesn’t look like any NAs based on info above.
- Overall, appears to be a legitimate trend in rising temperatures
Years Ranked by Average Daily Temp
Linear model of annual temperature change
##
## Call:
## lm(formula = mean.mean ~ ynum, data = vw.temp.yr)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.11837 -0.35274 -0.08088 0.36937 0.97701
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -40.987746 9.420792 -4.351 6.87e-05 ***
## ynum 0.025708 0.004722 5.444 1.66e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.4964 on 49 degrees of freedom
## Multiple R-squared: 0.3769, Adjusted R-squared: 0.3642
## F-statistic: 29.64 on 1 and 49 DF, p-value: 1.659e-06Model interpretation:
- Average daily temperature changing 0.0257081 degrees each yr.
- P-value for statistical significance: 1.6594794^{-6}
- Temperature change is statistically significant over this period
Look at Average Daily Maximum and Minimums for each Year
Linear modelling for max and min
##
## Call:
## lm(formula = mean.max ~ ynum, data = vw.temp.yr)
##
## Residuals:
## Min 1Q Median 3Q Max
## -0.93114 -0.31913 0.04793 0.35363 1.05197
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -27.048385 9.663318 -2.799 0.007310 **
## ynum 0.020496 0.004844 4.232 0.000101 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5092 on 49 degrees of freedom
## Multiple R-squared: 0.2676, Adjusted R-squared: 0.2527
## F-statistic: 17.91 on 1 and 49 DF, p-value: 0.0001014Model interpretation:
- Average daily MAX temperature changing 0.0204959 degrees each yr.
- P-value for statistical significance: 1.0137692^{-4}
- Temperature change is statistically significant over this period
##
## Call:
## lm(formula = mean.min ~ ynum, data = vw.temp.yr)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.30263 -0.31189 0.01655 0.40452 1.11209
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -54.893664 9.779701 -5.613 9.18e-07 ***
## ynum 0.030880 0.004902 6.300 8.07e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 0.5153 on 49 degrees of freedom
## Multiple R-squared: 0.4475, Adjusted R-squared: 0.4362
## F-statistic: 39.68 on 1 and 49 DF, p-value: 8.07e-08Model interpretation:
- Average daily MIN temperature changing 0.0308802 degrees each yr.
- P-value for statistical significance: 8.069551^{-8}
- Temperature change is statistically significant over this period
Monthly average temperature
Monthly Temperatures: Ranges
Monthly Temperature Trends over Years
Monthly Temp. Comparison YoY
For each month, what has been the pattern in ave. temperature over the years?
Seasonal Temperatures
Season definitions:
- Winter: Dec, Jan, Feb
- Spring: Mar, Apr, May
- Summer: Jun, Jul, Aug
- Fall: Sep, Oct, Nov
Note: the year for a winter season is applied to year at end of season. So winter from Dec 2018 to Feb 2019 is considered winter of 2019.